Moral Dimensions
Moral Dimensions PERMISSIBILITY, MEANING, BLAME T. M. Scanlon Belknap Press of Harvard University Press Cambridge, Massachusetts London, England 2008
Copyright 2008 by the President and Fellows of Harvard College All rights reserved Printed in the United States of America Library of Congress Cataloging-in-Publication Data Scanlon, Thomas. Moral dimensions : permissibility, meaning, blame / T. M. Scanlon. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-674-03178-4 (alk. paper) 1. Ends and means. 2. Intention. 3. Blame Moral and ethical aspects. I. Title. BJ84.E5S23 2008 170.42 dc22 2008016218
For Lucy
Contents Preface ix Introduction 1 1 The Illusory Appeal of Double Effect 8 2 The Significance of Intent 37 3 Means and Ends 89 4 Blame 122 Notes 217 Bibliography 239 Index 243
Preface This book has its origins in discussions at meetings of the Society for Ethical and Legal Philosophy. The sig nifi cance of the distinction between intended and merely foreseen consequences was a recurrent topic over the course of twenty years, and a subject of strong disagreement within the group. I myself was always torn. The distinction seemed to fit well with my judgments of right and wrong in many cases better than any other account I could think of. But I was unable to explain why this difference in an agent s attitudes should have such moral importance, and unable to defend the distinction against the counterexamples that Judith Thomson and others would raise. Further discussions with Judy and others in 1996,
Preface about whether we should rely on intent as an important factor in the argument of an amicus curiae brief concerning assisted suicide, in ten si fied my feeling of being intellectually whip sawed on this issue drawn one way by the apparent ability of double effect to explain particular cases, and in the opposite direction by my inability to defend it as a general principle. Hoping to come to some settled opinion on the question, I made it the topic of my graduate seminar at Harvard in the spring of 1999. I was fortunate to have an outstanding group of students in that seminar, who were willing to work through the extensive and diffi cult literature on double effect and on killing and letting die. My main objectives were to understand whether the permissibility of an action can depend on whether a given harm is intended by the agent or merely foreseen as a side effect of his or her action, and, if permissibility does not depend on intent in this way, to understand why the idea that it depends on intent should seem so plausible. By the end of the semester I had come up with provisional answers to these questions, which after many revisions became Chapter 1 of this book. My first debts of gratitude, then, are to my fellow members of the Society for Ethical and Legal Philosophy, for many years of stimulating discussions, and to the members of my Harvard seminar, for their critical acumen and their goodspirited willingness to tramp through jungles of argument in pursuit of an elusive quarry. The first version of Chapter 1 was presented to the Joint Session of the Aristotelian Society and the Mind Association in the summer of 2000, and published as Permissibility and x
Preface Intention I in Proceedings of the Aristotelian Society, Supp. Vol. 74 (2000): 301 317. Jonathan Dancy was my co-symposiast, and his comments helped me to see some of the ways in which that article could be improved. All four chapters of this book have been presented as lectures at many universities and have bene fited from the comments I received on those occasions. In particular, Chapters 1, 2, and 4 were discussed at the New York University Colloquium in Legal, Political and Social Philosophy. The first version of Chapter 3 was the Gareth Evans Memorial Lecture at Oxford, November 16, 2004. Chapter 4 was first presented as the Jack Smart Lecture at the Australian National University in Canberra, in July 2003; later versions were given as the Seybert Lectures at the University of Pennsylvania, and as the Howison Lecture at the University of California, Berkeley. I am grateful to the audiences on all these occasions for stimulating and helpful discussion. Derek Parfit, the most loyal friend and helpful critic one could imagine, read multiple versions of all of these chapters. His extensive comments have saved me from many errors and made the book far better than it otherwise would have been. The book has also bene fited from the comments of other friends, students, and colleagues who read or heard versions of these chapters. They include (at least) Harry Adamson, G. A. Cohen, Joshua Cohen, Tom Dougherty, Patricio Fernandez, Samuel Freeman, Jorge Garcia, Aaron Garrett, Kent Greenawalt, Paul Guyer, Pamela Hieronymi, Waheed Hussain, Aaron James, Frances Kamm, Erin Kelly, Michael Joel Kessler, Niko Kolodny, Jeff McMahan, Japa Pallikkathayil, Philip Pettit, Robert Post, Raffaele Rodogno, Samuel Scheffler, xi
Preface Tommie Shelby, Seana Shiffrin, John Skorupski, Angela Smith, Robert Stalnaker, Sigrun Svavarsdottir, Dennis Thompson, Judith Thomson, Alec Walen, R. Jay Wallace, Ralph Wedgwood, and Roger Wertheimer. I thank all of them for their helpful comments and suggestions. I am grateful, as always, to my wife, Lucy, for her understanding and support. xii
Moral Dimensions
Introduction The doctrine of double effect holds that an action that aims at the death of an innocent person, either as its end or as a means to its end, is always wrong. In particular, it holds that such an action cannot be jus ti fied by its good effects, such as saving the lives of a greater number of innocent people. This doctrine gains plausibility from its ability to explain some otherwise puzzling cases. For example, if the limited amount of a drug that is available could be used either to save one patient or to save five others, it is permissible to give it to the five, even though the one will die. But it would not be permissible to withhold the same drug from the same person in order to save the five others by transplanting his organs into them after he is dead. According to the doctrine of double effect, such an action would be impermissible because in the latter case, unlike
MORAL DIMENSIONS in the former, one withholds the drug from the one in order to bring about the death of the one, as a means to save the five. Similarly, many people believe that in war it can be permissible to bomb a military target even though this will also cause the deaths of some noncombatants living nearby, but that it would not be permissible to bomb the same number of noncombatants in order to hasten the end of the war by demoralizing the population. Again the doctrine of double effect seems to offer an explanation: in the latter case, but not the former, one would be aiming at the deaths of innocent people (people who pose no threat) as a means to one s end. The ability to explain cases of this kind gives the doctrine of double effect considerable initial appeal. But it is not at all clear how an agent s intentions could determine permissibility of an action in the way that the doctrine claims. To my knowledge, no convincing theoretical explanation for it has been offered, and arguments against this idea by Judith Thomson and others have seemed to me persuasive. If one rejects the doctrine of double effect, as I am inclined to do, then one is left with two questions. First, if this doctrine is not the correct explanation in cases like the ones I have mentioned, what is? Second, if the doctrine of double effect is mistaken, why should it seem so plausible as an explanation for these cases? What mistake underlies its appeal? I believe that the illusory appeal of double effect arises from a confusion between two closely related forms of moral judgment, which can be based on the same moral principles. Judgments of the first kind use a principle in what I call its deliberative employment to answer a question of permissibility, whether an agent may perform an action of a certain kind. 2
Introduction Judgments of the second kind employ principles in their critical use, to assess the way in which an agent went about deciding what to do on a given occasion. Judgments of the latter sort depend on what the agent saw as reasons bearing on the decision about what to do. Even when the permissibility of an action does not depend on what the agent took to be reasons, it may appear to do so if judgments of these two kinds are not clearly distinguished. I present this explanation in detail in Chapter 1. This explanation of the appeal of the doctrine of double effect depends on a particular understanding of the idea of moral permissibility, which leads to several further questions about how the disagreement between supporters and critics of the doctrine should be understood. This might be a disagreement about what determines the permissibility or impermissibility of certain actions. But the disagreement might arise instead from the fact that the parties are asking different basic questions. It might be that although many opponents of double effect are concerned with the question of permissibility, many supporters of double effect take some other moral notion, such as the idea of a good action, to be basic. If so, this raises further questions about what these different moral notions are and what grounds there are for assigning one of them rather than another a central role in our moral thinking. I raise these questions in Chapter 1, but I do not undertake to answer them, since the first question, in particular, is one that supporters of double effect are best placed to answer. I hope that my argument will lead to further discussion of this question. Nor do I offer a single direct answer to the first of the questions I posed above: if the doctrine of double effect is not 3
MORAL DIMENSIONS the correct explanation in cases like those I described at the outset, what is the correct explanation? My own view is that there is no single explanation for all of the cases to which the doctrine of double effect is thought to apply. The explanation of what is permissible or impermissible in transplant cases is different from what explains the difference between terror bombing and tactical bombing, and different still from the explanation of what is permissible when dealing with runaway trolleys. (I discuss these cases in Chapter 1 and Chapter 3.) If the permissibility of an action does not depend on the agent s intentions in the way that the doctrine of double effect maintains, this leaves open the possibility that there are other ways in which permissibility does depend on intent, or on an agent s reasons for acting. In Chapter 2 I identify a number of ways in which permissibility depends on an agent s intentions, and I examine some other ways in which it may appear to depend on an agent s intentions but does not do so, at least not in a fundamental way. In the course of this investigation I develop and explore an important distinction between the permissibility of an action and what I call its meaning the sig nifi cance, for the agent and others, of the agent s willingness to perform that action for the reasons he or she does. Although permissibility does not, in general, depend on an agent s reasons for action, meaning obviously does, and many of the cases in which permissibility depends on an agent s reasons for action are cases in which permissibility depends on meaning. In Chapter 3 I examine the ideas of treating a person as an end in him- or herself and treating a person as a mere means. Whether an action involves treating someone as an end or as a means depends on what the agent sees as reasons 4
Introduction for treating that person one way or another. So if the injunction to treat humanity always as an end in itself and never merely as a means is a criterion of permissibility, there is an important class of cases in which the permissibility of an action depends on what the agent sees as reasons bearing on that action. I investigate this question first by considering the Kantian idea of treating rational nature as an end in itself, taking the idea of treating rational nature as a means to be identical with failing to treat it as an end. Bearing in mind the distinction, introduced in Chapter 1, between two kinds of moral judgment, I argue that this Kantian idea can be understood in either of two ways. It is very plausible to say that an action is permissible only if it is consistent with the idea of rational nature as an end in itself. Whether a given action sat is fies this criterion, however, depends on the reasons for or against so acting, and not, in general, on what the agent takes those reasons to be. But the claim that, in a given action, an agent treated someone as an end, or failed to do so, can also be an observation about what the agent saw as reasons for acting one way rather than another. So understood, this is not a claim about the permissibility of the action but rather about its meaning. Distinct from the Kantian ideas of treating someone as an end in him- or herself or merely as a means, there is a more spe cific idea of treating someone as a means, or using them, that applies only in cases in which some effect on a person is causally necessary to the achievement of the agent s aims. This idea is sometimes invoked, not as a characterization of wrongful action in general, but as an explanation for why actions of a certain type are wrong. The fact that an action involves using 5
MORAL DIMENSIONS someone as a means in this more spe cific sense does sometimes make it fall within a broad class of wrongful action actions that are wrong because they involve others in ways that have certain costs to them, without their consent. But this class of wrongs can be explained without reference to the idea of using someone, or of acting in a way that causally depends on someone as a means. I argue that these ideas are not in themselves of basic moral sig nifi cance as determinants of permissibility. To say that in a certain action the agent was just using another person can, however, be an observation about the meaning of that action, and the fact that an action has this meaning can sometimes be relevant to its permissibility. To say that an action is blameworthy is to make a claim about its meaning: to claim that the action indicates something about the agent s attitudes that impairs his or her relations with others. To blame someone, in my view, is to understand one s relations with that person as modi fied in the way that such a judgment holds to be appropriate. In Chapter 4 I elaborate and defend this interpretation of blame, explaining how it differs from and should be preferred to other interpretations that take blame to be a kind of negative assessment, a sanction, or the expression of some moral emotion, such as resentment. I examine the implications of this interpretation for the ethics of blame for who may be blamed, who has standing to blame, and when one must blame. I also examine why blame might be thought to be appropriate only for actions that are undertaken freely, and explain why moral blame, as I understand it, does not presuppose free will. In the course of this book I argue for a number of particular moral claims, including claims about which actions are 6
Introduction permissible, about when intent matters to permissibility, and about various forms of moral responsibility. I hope that readers will be persuaded by what I have to say about these questions. But another important aim of the book is to identify and call attention to differences between the general moral notions in terms of which these particular judgments are expressed. These general notions are the dimensions of moral thinking referred to in the title: permissibility, meaning, and blame. My main claims are about how permissibility is to be understood; how it differs from meaning; how blameworthiness is a species of meaning; and how blame can be understood as a class of responses to this kind of meaning. I hope that readers who disagree with my particular moral claims will be led to re flect on the way that they understand these general moral notions. In particular, I hope they will consider whether the question of right and wrong as they understand it is what I am calling the question of permissibility. Similarly, I hope that those who disagree with me about blame will be inspired to make clear what they take moral blame to involve, and why blame as they understand it should be thought to require a particular kind of freedom. 7
1 The Illusory Appeal of Double Effect How does the moral assessment of an action, in particular the moral permissibility of performing it, depend on the agent s intentions? This chapter and the next undertake an in quiry into this question. My main aim in this chapter is to explain why intent seems to be relevant to permissibility in the particular way claimed by the doctrine of double effect, even though this apparent sig nifi cance is illusory. This leaves open the question, to be investigated in Chapter 2, of whether intent is relevant to permissibility in other ways. My second aim is to characterize the idea of moral permissibility and the ways in which it differs from other moral notions. I will begin with a few preliminary remarks about permissibility as I understand it.
The Illusory Appeal of Double Effect Permissibility The question of permissibility, May one do X? is generally asked as a precursor to a decision about whether to do X or not. So asked, it applies only to something that can be the object of an agent s decision, something that, if one were to decide to do it, would be done intentionally. But the question of permissibility, as I understand it, is broader than this. The question is not always a precursor to a decision but can also be asked retrospectively, or hypothetically, about the way a person behaved on some past, or imagined, occasion. One can ask, for example, whether it was permissible for a person to act in a way that posed a certain risk of harm to others, even if the person was unaware of that risk. The question of per missibility is always about something a person does or might do a way he or she might govern him- or herself and thus about something that could be the object of a possible decision. But it can be asked about a way of behaving characterized in a way that the agent was, at the time, unaware of, and hence about a way of behaving that was not intentional. Any action can be characterized in indefi nitely many ways. The phrase that replaces X in the question of permissibility, May one do X? will be only a partial characterization, mentioning only a few of the properties that a proposed action would have, usually ones that are thought to raise a question about its permissibility. Thus, for example, I might ask, May I give the patient this drug, given that it may well hasten his death? The correct response may of course be It depends, followed by a spec i fi ca tion of additional facts about the action 9
MORAL DIMENSIONS that are needed to determine its permissibility. Even this fuller specifi ca tion will not uniquely identify an action. Various factors that characterize an action, such as the exact time and the exact way in which the agent will move her body, are often irrelevant to permissibility. What I wish to investigate is when and why the features of an action on which its permissibility depends will include facts about the agent s intentions in performing it. Intent and Reasons Intention is commonly used in a wider and a narrower sense. 1 When we say that a person did something intentionally, one thing we may mean is simply that it was something that he or she was aware of doing or realized would be a consequence of his or her action. This is the sense of intentionally that is opposed to unintentionally : to say that you did something unintentionally is to claim that it was something you did not realize you were doing. But we also use intention in a narrower sense. To ask a person what her intention was in doing a certain thing is to ask her what her aim was in doing it, and what plan guided her action how she saw the action as promoting her objective. 2 To ask this is in part to ask what her reasons were for acting in such a way which of the various features of what she realized she was doing were features she took to count in favor of acting in this way. This narrower sense of intention is at least very close to the sense of intention involved in the distinction, central to the doctrine of double effect, between the consequences of one s action that are in- 10
The Illusory Appeal of Double Effect tended (as ends or chosen means) and those that are merely foreseen. One thing that these two senses of intention have in common is that each tells us something about an agent s view of the reasons bearing on his or her action. This is most obvious in the case of the narrower notion. Intention in the broader sense the idea of what one does intentionally is in the first instance a matter of what the agent understands herself to be doing rather than what her reasons were for doing it. But it is also true that if an agent does something intentionally in this broader sense if she is aware of a certain aspect of her situation then even if she does not take this aspect of what she is doing to provide a reason for her action, she at least does not (insofar as she is not acting irrationally) take it to constitute a suffi cient reason not to act in that way. In either case, then, claims about what an agent does intentionally, or intends, have implications about her assessment of the reasons that bear on acting in that way. Knowing an agent s intention in acting, and which things she is doing intentionally, can thus give us at least three kinds of information about the agent and her action. First, her intention, in the sense of her plan of action, tells us how she expects to move her body and to affect the world around her. Second, what it is that she is doing intentionally tells us what she believes about her situation and the likely effects of her action. Third, it also tells us something about how she evaluates these factors which she sees as reasons for acting the way she plans to act, which as costs to be avoided if possible, which as costs to be borne, which as inconsequential. 11
MORAL DIMENSIONS The Apparent Significance of Intent With this as background, let me turn to my main question, which is whether and how an agent s intention is relevant to the moral permissibility of what he or she does. This relevance may seem obvious. An agent s intention, it might be said, determines what the agent s action is whether it is an instance of lying, for example, or murder and what an action is in this sense must surely be relevant to its moral permissibility. There is an important way in which this is quite true but not directly relevant to the question I am pursuing. An agent s aim in acting, her plan in acting, her beliefs about the likely effects of her action, and her evaluation of various features of her situation which of them she sees as providing reasons, positive or negative, and which she sees as inconsequential are, taken together, an important part of our basis for predicting the effects that her action will have. They are not, however, a complete basis for predicting those effects. An agent s understanding of her situation is often incomplete or mistaken, and the effects of a planned action may be quite different from what the agent expected. So, to assess an action we may need to draw on other information. But the agent s intention, and what she is doing intentionally, are a crucial starting point. Moreover, the agent s evaluation of various features of her situation tells us how her action is likely to be guided which aspects of her situation she will attend to and how she is likely to respond if things do not go as she expects them to. These factors are all clearly relevant to the permissibility of an agent s action, and later I will consider their sig nifi cance in more detail. But this way in which intent can be relevant to 12
The Illusory Appeal of Double Effect the permissibility of an action is in an important sense derivative. What are of fundamental relevance in these cases are the effects of the agent s action on the world around her (or what it is reasonable to expect those effects to be). Her intention is relevant in the ways just described only because it tells us something about those effects. This is what I will call the predictive sig nifi cance of intent. But the question I am interested in is whether an agent s intention is itself directly relevant to the permissibility of an action whether, holding effects (or expected effects) constant, the permissibility of an action can depend on the agent s intention in performing it or, more generally, on what he or she saw as a reason for the action. So I will set aside for the moment the predictive sig nificance of intent and concentrate on the question of how intent might be more directly relevant to permissibility. Various lines of thought support the conclusion that intent is directly relevant to the moral assessment of an action. There is obviously an important moral difference between intentionally harming someone, causing harm negligently, and doing so through a freak accident. The difference between causing harm intentionally and doing so negligently, however, is not a difference in permissibility. Both are generally impermissible. The difference between them lies, rather, in the kind of fault that is involved when an agent acts impermissibly in these ways. The difference between causing harm in either of these ways and causing harm through a freak accident does appear to be a difference in permissibility. But what makes the difference here is not intent. What differentiates negligence from a freak accident is not the agent s aims, or necessarily 13
MORAL DIMENSIONS what the agent believed about the likely effects of his or her action, but what he or she should have believed, under the circumstances, about the likely effects of that action. It has seemed to many people that there are cases in which not just the overall moral assessment of an action but also its permissibility depends on the agent s intention. In wartime, for example, as I mentioned in the Introduction, it seems that there is an important moral difference between attacking a military target in a way that can be foreseen to lead to a certain number of civilian casualties and killing the same number of civilians in order to demoralize the population or discourage them from aiding the enemy. The difference between these two lines of action may seem to be one of intent. It may seem to lie in the fact that in the latter case, but not the former, those who carry out the attack intend to kill the civilians. They are not merely aware that this will be the effect of what they are doing; they are aiming at it, as a means to their end. More generally, many people have been drawn to the doctrine of double effect, which holds that, although it can be permissible to do something that one can foresee will lead to the deaths of innocent people, when doing it is necessary to achieve some greater good, it is impermissible to kill the same number of innocent people as a means to achieving the same greater good. Although this doctrine is controversial, it is appealing because it seems to offer the best explanation of the distinction between terror bombing and tactical bombing, and also to explain other cases, such as the following. 14 Drug Shortage: There are five people in Room B, and one person in Room A, all of whom have the same disease,
The Illusory Appeal of Double Effect and all of whom will die if not treated soon. There is enough medicine on hand to cure all five of the people in Room B, but since the person in Room A has a more advanced case, it would take all of the available supply to save just him. In this case it is clearly at least permissible to use all of the medicine to save the people in Room B. Now consider a different situation. Transplant: The five people in Room B are in need of organs one needs a heart, two need a lung, and two need a kidney and they will all die if they are not given transplants soon. Unfortunately, no organs are presently available. But there is a person in Room A, in for a checkup, who could be given a lethal injection instead of the inoculation he is expecting, thereby making his organs available to save the five others. This is clearly impermissible. It might be tempting to explain the difference between these cases by saying that although it is permissible to let one die in order to save five, as in Drug Shortage, what is proposed in Transplant is to kill the one in order to save the five, and this is not permissible. Drug/Transplant: The people in Room B are the same as in Transplant, and the person in Room A is the same as in Drug Shortage. If the person in Room A dies of his illness, his organs will not be damaged and thus can be used to save the five. Is it permissible to withhold the drug from him? 15
MORAL DIMENSIONS Everyone with whom I have discussed these cases agrees that it is not. Why not? One answer would be that, as in Transplant, this would be a case of killing one to save five. But this cannot be the basic explanation. What is proposed in this case is the same as in Drug Shortage to refrain from giving the person in Room A the available drug and it was said in that first case that this would not be killing. If it is killing in this case, then this is because killing is being used as a moral notion that is to say, it is killing because it is wrong (for some other reason) rather than being wrong because it is killing. What, then, is this other reason? The doctrine of double effect offers an explanation. Whereas in Drug Shortage the death of the one is merely a foreseeable consequence of giving the drug only to the five, in Drug/Transplant the death of the one is intended. What is proposed in that case is to withhold the drug from the one precisely in order to bring about the patient s death right away, so that his organs will be available for transplant in time to save the five. It is thus a case of intentionally killing an innocent person, which is absolutely impermissible according to the doctrine of double effect. It might be suggested that the wrong involved in cases like Drug/Transplant is to be explained by rules that are peculiar to the hospital setting rules specifying the duties that hospital personnel owe to patients rather than by general moral principles. (Just as it might seem that the wrong involved in terror bombing is explained by the laws of war.) I will return to this possibility below, but it is worth noting here that what seem to be similar wrongs can occur in cases that do not 16
The Illusory Appeal of Double Effect involve hospitals or the obligations of doctors and nurses. Consider the following three cases. Rescue I: As I am driving home, I hear on my citizensband radio that a car is stalled along a seldom-traveled road that I could easily take. The driver of the car is trying to deliver medicine to someone who will die unless he receives it within the next few hours. I could easily take that road and restart the stalled car. Clearly I should do so. Rescue II: Same as the previous case, except I also hear that along another road I could take there is a stalled car that was taking medicine to five people in equally urgent need. There is not enough time for me to go to the aid of both cars. Clearly in this case it is at least permissible for me to aid the second car, so as to save five rather than only one. Rescue/Transplant: Same as Rescue I, except that I know that there are five people in urgent need of transplants who will be saved if the patient awaiting the medicine dies very soon, as he will if I do not go to the aid of the stalled car. It seems clearly impermissible to refrain from aiding the car in this case. As before, one plausible explanation is that in this case, but not in Rescue II, I would be intending that the one person should die, as a means to saving the five. Like many others, I have found this explanation appeal- 17
MORAL DIMENSIONS ing. But there are well-known problems with it. First, to my knowledge no one has come up with a satisfying theoretical explanation of why the fact of intention, in the sense that is involved here the difference between consequences that are intended and those that are merely foreseen should make a moral difference. Second, there are cases in which applying this distinction seems to give the wrong answer. For example, in the well-known trolley-problem case it seems permissible to switch a runaway trolley onto a sidetrack on which it will hit only one person rather than allow it to continue straight ahead and hit five. But it also seems permissible to switch the trolley in the Loop case, proposed by Judith Thomson, in which the sidetrack loops around and rejoins the main line, so that if the trolley does not hit the one person and thereby come to a stop, it will continue around the loop and hit the five from the other side. 3 The answer in this case may be less clear than in the original one, but it is at least quite plausible to maintain that if it is permissible to turn the trolley onto the sidetrack in the first case, then it is permissible to do this in the second case as well. These cases seem to differ in just the way I have described: in the second, but not the first, one switches the trolley only because it will, by hitting the one person, be prevented from hitting the five. This person s being hit by the trolley is intended as a means to the end of saving the others. Perhaps the distinction between harms that are intended and those that are merely foreseen, even though it makes a moral difference in the drug and transplant cases described above, does not make a difference in these cases. But this needs to be explained. 4 18
The Illusory Appeal of Double Effect Thomson also presents, as counterexamples to the kind of analysis we are considering, cases involving the use of lethal drugs for pain relief. Suppose that a patient is fatally ill and in great pain. The only course of medication that will relieve this pain will also cause the patient s death. Suppose that the patient wants to be given this drug. Does the permissibility of administering it depend on the doctor s intention in doing so spe cifi cally, on whether the doctor intends to relieve the pain by causing the patient to die or intends to relieve the pain by giving the drug, which will, inevitably, also cause the patient s death? Thomson says, plausibly, that it does not. This conclusion may draw support from the thought that it is not a bad thing, morally speaking, for a person in such circumstances to die sooner rather than later, and that therefore the usual moral strictures against causing death do not apply. But Thomson s objection retains its intuitive force in cases in which death is clearly a bad thing, such as military cases like the one I mentioned earlier. It is plausible to claim that it can be permissible in wartime to bomb a munitions factory even though this is certain to kill some civilians living nearby, but that it would not be permissible to kill the same number of civilians just as a way of undermining public support for the war, even if doing this would hasten the end of the conflict just as much as destroying the munitions plant would. These two actions clearly differ in moral permissibility. But is it equally clear that the moral difference between them is a matter of what is intended by the agents involved? Thomson says that it is not. This would seem to be supported by thought experiments such as this: Suppose you were prime minister, and the commander of the air force described 19
MORAL DIMENSIONS to you a planned air raid that would be expected to destroy a munitions plant and also kill a certain number of civilians, thereby probably undermining public support for the war. If he asked whether you thought this was morally permissible, you would not say, Well, that depends on what your intentions would be in carrying it out. Would you be intending to kill the civilians, or would their deaths be merely an unintended but foreseeable (albeit ben e fi cial) side effect of the destruction of the plant? 5 Holding fixed the ac tual consequences of the raid and what the parties have reason to believe these consequences to be, might an action be permissible if performed by an agent with one intention but impermissible if performed by an agent with a different strategy in mind? I agree with Thomson in find ing this implausible. Explaining the Appeal of Double Effect If this is implausible, why should the doctrine of double effect have seemed appealing as an explanation in the cases I considered earlier? Thomson suggests that the appeal of that doctrine depends on a failure to take seriously enough the fact I think it is plainly a fact that the question whether it is morally permissible for a person to do a thing is just not the same as the question whether the person who does it is thereby shown to be a bad person. 6 As she says, a doctor who dislikes her patient and administers a lethal dose of pain killer, relishing the thought that this will be the last of him, is moved by morally objectionable reasons, even if an earlier death is in fact better for her patient. If she is moved by such reasons, then she is a morally bad person. But it does not follow that it is im- 20
The Illusory Appeal of Double Effect permissible for her to administer the drug (or that the patient should have to wait until a different doctor, with better intentions, comes on duty). There is something right in the suggestion that the appeal of the doctrine of double effect arises from a failure to distinguish clearly between assessing an agent and assessing the permissibility of her action. But in the form in which Thomson states it, this does not seem to explain the apparent sig nifi cance of an agent s intention in cases of the kind we have been considering. Malicious thoughts of the kind just described, whether or not they operate as motives, certainly reflect badly on the character of the doctor. But no such thoughts have been involved in the cases we have been considering. The doctors in Transplant and Drug/Transplant are, we assume, moved purely by the desire to save as many lives as possible. The fact that they aim to do this by sacrificing an innocent person does not show them to be of bad character unless this is, in de pen dently, something that it is wrong to do. It therefore does not seem that we are tempted to think that it is wrong because it clearly indicates bad character and we fail to distinguish between the assessment of character and the question of permissibility. 7 I believe that the truth in Thomson s point can be captured by put ting it in a different way. The illusory appeal of intent as a way of explaining cases like Drug/Transplant flows, I suggest, from two facts about moral principles. The first is that although principles stating moral requirements specify that certain considerations normally count decisively for or against acting in a certain way, these principles almost always allow for exceptions. So, for example, according to the principle of 21
MORAL DIMENSIONS fidelity to promises, the fact that one promised to do a certain thing is normally a conclusive reason for doing it a reason that determines what one ought to do even if it would be more convenient or more advantageous to do something else. But there are exceptions. For example, one need not, and indeed should not, fulfill a promise to one person to do something fairly trivial if doing so would cause great harm to someone else. Fully understanding the morality of promising involves being able to recognize the considerations that do, and those that do not, justify such exceptions. The second feature of moral principles that I have in mind is that they can be employed in either of two ways: as standards of criticism or as guides to deliberation. As guides to deliberation, moral principles answer a question of permissibility: May one do X? They also explain the answer by identifying the considerations that make it permissible or impermissible to do X under the circumstances in question. These considerations may concern what the agent sees as reasons for acting, or other features of his state of mind, but they need not and often do not do so. In what I will call their critical employment, however, a principle is used as the basis for assessing the way in which an agent went about deciding what to do on some real or imagined occasion. Used in this way, it provides the basis for answering a question of the form, In deciding to do X under those circumstances, did Jones take the proper considerations into account and give them the right weight? An answer to this question depends on an answer to the prior question, of which considerations are relevant to the permissibility of such an action and how they should be taken into account. But it goes beyond that question in asking whether 22
The Illusory Appeal of Double Effect the agent in question in fact took those considerations into account in the proper way. 8 These two uses of moral principles are closely related, but there is a crucial difference between them. Criticism of the way an agent decided what to do is unavoidably predicated on assumptions about the agent s state of mind in particular about what he or she took into account in deciding what to do and took as reasons for and against acting as he or she did. By contrast, when principles are used to guide deliberation, they do this merely by specifying which considerations do, and which do not, count for or against various courses of action. I will refer to these applications of principles as their critical and their deliberative uses. 9 It is easy to overlook this distinction. Since principles tell agents which considerations count for or against an action, it is natural to say that agents follow these principles when they take these considerations as reasons, and that when they do not, their failure to do so makes their actions wrong. But what makes an action wrong is the consideration or considerations that count decisively against it, not the agent s failure to give these considerations the proper weight. Suppose, for example, that I have promised to sell you my house, and that under the circumstances this counts as a decisive reason for doing so. In particular, the fact that I could get more money by breaking my promise and selling the house to someone else is not a suffi cient reason to do that. But suppose I do break the promise in order to get this bene fit. In describing what was defective about my action, you might say that I acted wrongly in taking my own advantage as suffi cient reason to break my promise. This would be true, as a critical 23
MORAL DIMENSIONS observation about the decision I made. At a more fundamental level, however, what made my action wrong was not the fact that I acted for a bad (selfish) reason, but rather the fact that I had promised to sell you the house. Given that there were no countervailing considerations that would justify an exception to the requirement that promises be kept, this fact counted decisively against my selling the house to someone else. The distinction is even clearer when we view the case prospectively. Suppose I ask, while deciding what to do, Must I do what I promised? Why shouldn t I sell to the other person, since it would be more profit able? You would not reply, That would be wrong, because you would be aiming at your own benefit (or acting for the sake of your bene fit). It is true that I would be open to criticism for taking my own advantage to be suffi cient reason for breaking a promise. But this is criticism of the way I went about deciding, not an explanation of why my action would be wrong. What makes it wrong for me to sell to the second potential buyer is that I promised to sell to the first one. Supporters of the doctrine of double effect may disagree, or at least seem to disagree, because they take the relevant category of moral criticism to be that of a good action. Since an agent s intention in acting is crucial in making what he does an action, and making that action the type of action it is, they might argue, this intention is crucial in determining whether it is a good action. 10 Since they believe that an agent, in deciding what to do, should be guided by whether a proposed action is good, it might seem that they are proposing an alternative answer to the question of permissibility, and disagreeing with me 24
The Illusory Appeal of Double Effect about the relevance of intent in answering this question. What I am saying, however, is that intent seems relevant to permissibility only if one does not distinguish what I am calling the deliberative and critical uses of a principle. And indeed, the idea of a good action is naturally understood as falling on the critical side of my distinction. This raises the interesting question of whether those who disagree about double effect are disagreeing about the application of a single moral concept about what makes certain actions impermissible or whether they are employing different categories of moral appraisal one side talking about permissibility, the other about the goodness of actions. If they are employing different categories, this leads to the further question of what reasons there are for taking one or the other of these categories as the one that is important for certain purposes for example, as having the fundamental role in deliberation. I believe that the two features of moral principles that I have described provide the best explanation of our reactions to Transplant and to the other examples I have discussed above. First, although the examples differ in many ways, they all have the same structure: they concern general principles that sometimes admit of exceptions, and they raise questions about when those exceptions apply. For example, it is normally impermissible to act in ways that can be foreseen to cause serious harm. The runaway trolley case shows that this principle has exceptions: it is sometimes permissible to act in ways that can be foreseen to cause serious harm to some people if this is the only way to prevent similar harm to a greater number. Transplant shows that although the possibility of saving others sometimes justifies an exception to the prohibition against 25
MORAL DIMENSIONS causing foreseeable harm, it does not always do so. Similarly, Drug Shortage and Rescue II show that there are exceptions to the principle requiring us to aid others when we can: in these cases it is permissible to fail to save some because this is the only way to save others. But our reactions to Drug/Transplant and Rescue/Transplant re flect our judgment that the possibility of saving others does not always justify an exception to this principle. The underlying question raised by these cases is why the principles involved in them should have this particular form: why a consideration that justifies an exception to a principle in some cases should not do so in others. I will return to this question below. What I want to point out here is that if these cases have the structure I have just described, then the second feature of moral principles that I discussed above the distinction between the critical and deliberative uses of a principle, and the ease of overlooking this distinction can explain why it is tempting, but nonetheless mistaken, to think that they are cases in which the intentions of the agent play a fundamental role in determining permissibility. It is tempting to say that what would make it wrong for an agent to fail to save the person in Drug/Transplant or in Rescue/Transplant is the fact that she would be intending that person s death that is to say, she would be taking the advantages of his dying sooner as suffi cient reason not to save him. It would be quite correct to say this if we were taking the relevant principle in its critical employment and assessing the way in which the agent decided what to do. That is, it would be quite correct to say that her way of making this decision was defective because she took a certain consideration (the possi- 26
The Illusory Appeal of Double Effect bility of saving five others) as justifying an exception to the principle requiring one to aid others, when in fact that consideration did not justify an exception to the principle in this case. But the question of permissibility is answered by considering the relevant principles in their deliberative employment. What makes a proposed action wrong is the consideration that the relevant principle iden ti fies as counting decisively against it (given the absence of relevant countervailing considerations). In the promise-keeping example discussed above, this was the fact that I had promised to sell you my house. In Drug/ Transplant and Rescue/Transplant, it is the fact that there is a person who is in need of aid that the agent can easily provide. A person who takes the possibility of saving others via transplant as justifying an exception to the duty to aid in these cases, or takes the possibility of selling his house for a higher price as justifying an exception to the obligation to keep a promise, makes a mistake in so doing. What makes this a mistake, and what makes the corresponding actions impermissible, are the considerations that support the relevant obligations. The distinction in question here between the deliberative use of a principle to decide whether a certain action is permissible and its critical use to assess an agent s pro cess of decision making is similar to the distinction mentioned by Thomson, between assessing an action and assessing an agent. But the distinction I am calling attention to is narrower and, for that reason, easier to overlook. It is the distinction between the permissibility of an action and a special kind of agent assessment, in which what is being assessed is not the agent s 27